Learning Forum

[gzhang24]

Hello everyone,

I am working on thermal management of electrical machines. I need to use material with different thermal conductivity in different direction to represent the machine winding and somehow we find a way to define an anisotropic material using following code in command editor of material.

   THERMAL CONDUCTIVITY: 

    Option = Orthotropic Cartesian Components

    Thermal Conductivity X Component = 0.87 [W m^-1 K^-1]

    Thermal Conductivity Y Component = 0.87 [W m^-1 K^-1]

    Thermal Conductivity Z Component = 265.0 [W m^-1 K^-1]

The defined material works fine in single region. However, when two regions are set to be anisotropic material and they are using differnt coordinates, the interface of these two regions seems blocking heat transfer. The second picture is when I put a heat source point in the roughly middle of end winding and the heat cannot transfer to other regions.

I have tried changing factors in intersection settinng of mesh connection but it does not work out.

I wonder what can I do to make it works. Thanks.

[rfblumen]

If I understand correctly from your explanation of the set up, you have two solid domains, one embedded inside the other. Both solid domains use the same material that has orthotropic thermal conductivity. A point energy source is located in the inner solid. One of the domains uses a local coordinate system that has coordinate directions different from the global coordinate system. You're seeing a discontinuity in temperature at the interface between the two solids.
I set up an example case based on the above, however I wasn't able to see any temperature discontinuity at the interface. I then modified the example such that both solids used the global coordinate system, but had each solid use a different material. The material properties of the two materials were the same except of the direction values of the thermal conductivity. I then saw what appeared to be a "discontinuity" in temperature across the interface due to the difference in thermal conductivity across the interface. However, if the mesh at the interface is refined with appropriate inflation on both sides, the temperature is actually continuous across the interface but has a high gradient proximal to the interface.
You might try the same approach with your case and add mesh inflation to the two sides of the interface.

[gzhang24]

Thanks fot answering my question and your understanding is totally correct.
I just tried building up a easier model with only the solid winding but no fluid in CFX. However, I am having a more serious problem that the residual became so high. I think I must made some stupid mistakes but I couldn't find out.
Although this model must be wrong, it still shows the same discontinuity in temperature issue.
What confuses me most is that the interface of domin 1 and domin 2 looks fine with a discontnuous probably due to high residual, but the interface of domin 2 and 3 shows a very strange temperature distribution. The 3 domins are using same material that has orthotropic thermal conductivity, but they are using differnet coordinates.
Please give me some suggestions. Many thanks.